updating of calc/meas/proc/.count/intensity definitions

[rowlesmr]

Was initially looking just to address #42 .

But I now have other questions:

• pd_meas.counts/intensity_ should be Recorded, not Derived - Addressed in this initial commit.
• Some ambiguity in definitions of _pd_proc.intensity_norm and _pd_proc.intensity_net: which one contains a bkg? From the definition of _pd_calc.intensity_net, I'm guessing _pd_proc.intensity_net is background-corrected and normalisation-corrected, and the other is just normalisation-corrected.
• Should all _pd_proc.intensity_* and _pd_calc.intensity_* have their values limited to be > 0 (as is currently)? ( c.f. discussion in cif_core)
• The definitions of _pd_meas.counts/intensity_background/container/monitor say (amongst other things...)

  • Corrections for background, detector dead time etc. should not have been made to these values. Instead use _pd_proc.intensity_* for corrected diffractograms.

  • There is no direct _pd_proc. equivalent to these data names. Does this above paragraph mean that you should use _pd_proc.intensity_net/norm/total as appropriate, after doing the corrections? or are there missing _pd_proc.intensity_background/container/monitor data names?

[jamesrhester]

I like the changes made so far in this pull request.

_pd_calc.intensity_net and _pd_proc.intensity_net should certainly be calculated in the same way in order to be directly comparable, so I agree that they should be both normalised and background-subtracted.

I think restricting _pd_calc.intensity_net values to be >= zero is defensible. I would leave the other one unrestricted in recognition of the fact that current software does not always treat values close to zero in a way that would avoid small negative values, which would involve a bit of not-very-advanced Bayesian footwork. I am open to be convinced otherwise and don't really care either way.

The _pd_proc equivalents are missing because, as suggested, these _pd_meas data names have been used to calculate the total/net/norm intensities which the refinement works on. I don't see any need to add the equivalents.

If everyone is happy with these replies I think we can merge these changes.

[rowlesmr]

I've removed the enumeration range on _pd_proc.intensity_net, added some words in the description of _pd_proc.intensity_background/container/monitor talking about _pd_proc.intensity_net/norm/total, and clarified the bkg/norm correction for _pd_proc.intensity_norm/net.

I think that covered everything.